OPEN THE CASE FILE
X is arguing about kindness to AI. Is it a calculator? Could it feel anything? Does the way we talk to it say something about us?
This is a six-question horror-comedy test about the behaviour you choose. It doesn’t settle AI consciousness, predict dangerousness, or diagnose anyone.
Creator Daniel Hendricks reported swearing in 14.7% of his analysed submissions, but explicit frustration cues in 5.8%. “Fuck yes” can be praise. Profanity, frustration and cruelty are different things.
Those are one person’s chat-history figures, with a different method and denominator from this quiz. They are the inspiration, not a population benchmark or a menace score.
Each answer has an editorial weight from 0–3. Add the six weights, divide by 18, then round to a percentage. Neutral commands, critical feedback, happy swearing and testing a system score zero. Insults and threats add points. The labels and cutoffs are a joke, not a validated assessment.
0–16: Probably fine · 17–38: A little mean · 39–66: Menace to society · 67–100: You need help—with your prompts.
We never ask for your chats, name or account. Answers stay in this tab and aren’t sent to our server. A shared link contains only the score. Hosting services may keep ordinary access logs.
Independent satire. Not affiliated with any AI provider. Sound is optional. Reduced-motion preferences are respected.